Publication number | US20050007267 A1 |

Publication type | Application |

Application number | US 10/865,275 |

Publication date | Jan 13, 2005 |

Filing date | Jun 10, 2004 |

Priority date | Jul 9, 2003 |

Also published as | US6992606 |

Publication number | 10865275, 865275, US 2005/0007267 A1, US 2005/007267 A1, US 20050007267 A1, US 20050007267A1, US 2005007267 A1, US 2005007267A1, US-A1-20050007267, US-A1-2005007267, US2005/0007267A1, US2005/007267A1, US20050007267 A1, US20050007267A1, US2005007267 A1, US2005007267A1 |

Inventors | Thomas Zogakis, Himamshu Khasnis, Baireddy VIJAYAVARDHAN |

Original Assignee | Zogakis Thomas Nicholas, Khasnis Himamshu G., Vijayavardhan Baireddy |

Export Citation | BiBTeX, EndNote, RefMan |

Patent Citations (11), Referenced by (17), Classifications (10), Legal Events (3) | |

External Links: USPTO, USPTO Assignment, Espacenet | |

US 20050007267 A1

Abstract

System and method for a multi-standard sigma-delta modulator. A preferred embodiment comprises an integrator coupled to a signal input, wherein the integrator is configured to sum a difference between an input signal and an output of the sigma-delta modulator. A dithering circuit, also coupled to the signal input, is used to produce a dithering sequence. A summing point combines an output of the integrator and the dithering circuit while a quantizer converts the combined outputs into one of several discrete levels. The sigma-delta modulator further comprises a reset circuit to reset the sigma-delta modulator if the integrator becomes saturated. The sigma-delta modulator, as described, is compatible with G.Lite, ADSL, and ADSL+ variants of the digital subscriber line technical standards.

Claims(30)

an integrator coupled to a signal input, the integrator configured to sum a difference between an input signal provided at the signal input and an output of the sigma-delta modulator;

a dithering circuit coupled to the signal input, the dithering circuit configured to produce a dithering sequence;

a summing point coupled to the integrator and the dithering circuit, the summing point to combine outputs provided by the integrator and the dithering circuit;

a quantizer coupled to the summing point, the quantizer to convert an output of the summing point into one of several discrete levels; and

a reset circuit coupled to the integrator, the reset circuit configured to reset the sigma-delta modulator if the integrator becomes saturated.

a second summing point coupled to an input, the second summing point configured to combine a first scaled feedback signal from an output of the second-order integrator, a signal provided by the input, and a second scaled feedback signal from an output of the quantizer;

a first first-order integrator coupled to the second summing point, the first first-order integrator configured to accumulate an output produced by the second summing point;

a third summing point coupled to an output of the first first-order integrator, the third summing point configured to combine a third scaled feedback signal from an output of the quantizer and a signal provided by the first first-order integrator; and

a second first-order integrator coupled to the third summing point, the second first-order integrator configured to accumulate an output produced by the third summing point.

a magnitude quantization unit coupled to the signal input, the magnitude quantization unit configured to scale the amount of dithering based upon the magnitude of the input signal;

a dither generator configured to generate a dithering sequence based upon a pseudo-random number (PN) sequence; and

a fourth summing point coupled to the magnitude quantization unit and the dither generator, the fourth summing point to combine the dithering sequence with a scaled version of the input signal.

a magnitude unit configured to provide an absolute value of the input signal; and

a quantizer coupled to the magnitude unit, the quantizer to convert an output of the magnitude unit into one of several discrete levels.

an inverse Fourier transform unit coupled to a signal input, the inverse Fourier transform unit configured to convert a frequency domain symbol into a time domain symbol;

an interpolation chain coupled to the inverse Fourier transform unit, the interpolation chain configured to upsample the time domain symbol; and

a sigma-delta modulator coupled to the interpolation chain, the sigma-delta modulator configured to convert the time domain symbol into an analog signal, wherein the sigma-delta modulator includes a dithering circuit and a reset circuit.

an integrator coupled to the interpolation chain, the integrator configured to sum a difference between an input signal provided at the interpolation chain and an output of the sigma-delta modulator;

a dithering circuit coupled to the interpolation chain, the dithering circuit configured to produce a dithering sequence;

a summing point coupled to the integrator and the dithering circuit, the summing point to combine outputs provided by the integrator and the dithering circuit;

a quantizer coupled to the summing point, the quantizer to convert an output of the summing point into one of several discrete levels; and

a reset circuit coupled to the integrator, the reset circuit configured to reset the sigma-delta modulator if the integrator becomes saturated.

determining if an integrator in the sigma-delta modulator is operating in saturation;

resetting the sigma-delta modulator if the integrator has been operating in saturation for a specified amount of time.

examining the two most significant bits of a signal produced by an integrator in the sigma-delta modulator;

incrementing a counter if the two most significant bits are different; and

resetting the counter if the two most significant bits are equal.

combining a dithering sequence with an output of an integrator, wherein the integrator is a part of the sigma-delta modulator;

determining if the sigma-delta modulator has become unstable; and

resetting the sigma-delta modulator if the sigma-delta modulator has become unstable.

examining the two most significant bits of a signal produced by the integrator;

incrementing a counter if the two most significant bits are different; and

resetting the counter if the two most significant bits are equal.

Description

- [0001]This application claims the benefit of U.S. Provisional Application No. 60/485,939, filed Ser. No. 07/09/2003, entitled “Sigma-Delta Modulator for Multiple DSL Standards”, which application is hereby incorporated herein by reference.
- [0002]The present invention relates generally to a system for digital communications, and more particularly to a system and method for a multi-standard sigma-delta modulator.
- [0003]Digital-to-analog converters (DACs) can be used to convert digital signals into an analog representation of the same signal. DACs can be used in wide variety of applications, ranging from medical and entertainment to communications (both for voice and data). One form of DAC that is commonly used is what is known as a sigma-delta DAC. A sigma-delta DAC can make use of a sigma-delta modulator to convert an oversampled, band-limited digital signal into an output signal. Note that the output signal may be a single-bit (representing two levels) signal or a multi-bit (representing 2
^{N }levels, wherein N is the number of bits) signal. The output signal can then be converted into an analog signal via a conventional DAC and then filtered by a smoothing filter to provide an analog representation of the digital signal. An advantage of a sigma-delta DAC is the noise shaping that is inherent in the design of the DAC. The noise shaping increases as the oversampling rate of the sigma-delta DAC increases. It may be preferred that a sigma-delta modulator for a sigma-delta DAC oversample at a rate of 32 or 64 (or higher). - [0004]A design of a sigma-delta modulator can feature differing values for order (which can specify the sharpness of the noise shaping performed by the sigma-delta modulator), the number of bits in the output signal (which can affect the stability and noise level of the modulator), stability handling techniques, and so forth. A fairly standard sigma-delta modulator may be a first order sigma-delta modulator with a single-bit output signal. The design of such a sigma-delta modulator can be relatively straight forward, therefore, easy and inexpensive to design. Furthermore, with a single-bit output signal, the routing of the output signal line can be relatively simple.
- [0005]One disadvantage of the prior art is that of being a low-order sigma-delta modulator, noise suppression may not occur as rapidly as desired or needed, i.e., the order is not high enough. Sufficient noise suppression may simply require a higher order sigma-delta modulator, which can be more difficult to design and may not be able to operate at sufficient frequencies.
- [0006]A second disadvantage of the prior art is that the use of a single bit output line can require that the sigma-delta modulator operate at a higher frequency than a similar sigma-delta modulator with multiple bit output lines. Since power consumption rises with the square of the operating frequency, the sigma-delta demodulator operating at a lower frequency can consume less power.
- [0007]Another disadvantage of the prior art is that the high oversampling rate of the sigma-delta modulator (32, 64, or higher) can result in a design that may be hard to clock, especially if the signal being converted has a high clock rate of its own. For example, given a signal with a clock rate of 4 MHz, a sigma-delta modulator with an oversampling rate of 32 would require a clock of 128 MHz, while a 64 times oversampling sigma-delta modulator would require a clock of 256 MHz.
- [0008]These and other problems are generally solved or circumvented, and technical advantages are generally achieved, by preferred embodiments of the present invention which provides a system for a multi-standard sigma-delta modulator.
- [0009]In accordance with a preferred embodiment of the present invention, a sigma-delta modulator comprising an integrator coupled to a signal input. The integrator is configured to sum a difference between an input signal provided at the signal input and an output produced by the sigma-delta modulator. A dithering circuit is also coupled to the signal input. The dithering circuit is configured to produce a dithering sequence. A summing point combines outputs produced by the integrator and the dithering circuit, and a quantizer converts an output of the summing point into one of several discrete levels. The sigma-delta modulator also comprises a reset circuit coupled to the integrator. The reset circuit is configured to reset the sigma-delta modulator if the integrator becomes saturated. Other embodiments of the invention provide other features.
- [0010]In accordance with another preferred embodiment of the present invention, a multi-standard digital subscriber line (DSL) sigma-delta digital-to-analog converter (DAC) comprising an inverse Fourier transform unit that is coupled to a signal input, the inverse Fourier transform unit configured to convert a frequency domain symbol into a time domain symbol. An interpolation chain is coupled to the inverse Fourier transform unit and is configured to upsample time domain symbols. The multi-standard DSL sigma-delta DAC also comprises a sigma-delta modulator configured to convert the time domain symbol into an analog signal, wherein the sigma-delta modulator includes a dithering circuit and a reset circuit. Other embodiments of the invention provide other features.
- [0011]In accordance with another preferred embodiment of the present invention, a method for resetting a sigma-delta modulator comprising determining if an integrator in the sigma-delta modulator is operating in saturation. Furthermore, the sigma-delta modulator is reset if the integrator has been operating in saturation for a specified amount of time. Other embodiments of the invention provide other features.
- [0012]In accordance with yet another preferred embodiment of the present invention, a method for maintaining sigma-delta modulator stability comprising combining a dithering sequence with an output of an integrator, wherein the integrator is a part of the sigma-delta modulator. Additionally, a determination is made as to whether the sigma-delta modulator has become unstable. The method also comprises resetting the sigma-delta modulator if the sigma-delta modulator has become unstable. Other embodiments of the invention provide other features.
- [0013]An advantage of a preferred embodiment of the present invention is that with a 4th-order sigma-delta modulator, a good compromise can be achieved between noise shaping properties and design complexity.
- [0014]A further advantage of a preferred embodiment of the present invention is with dynamic dithering, the sigma-delta modulator's stability range can be increased.
- [0015]Yet another advantage of a preferred embodiment of the present invention is the use of a multi-bit output signal line permits a lower operating frequency for the sigma-delta modulator, thereby reducing power consumption.
- [0016]The foregoing has outlined rather broadly the features and technical advantages of the present invention in order that the detailed description of the invention that follows may be better understood. Additional features and advantages of the invention will be described hereinafter which form the subject of the claims of the invention. It should be appreciated by those skilled in the art that the conception and specific embodiments disclosed may be readily utilized as a basis for modifying or designing other structures or processes for carrying out the same purposes of the present invention. It should also be realized by those skilled in the art that such equivalent constructions do not depart from the spirit and scope of the invention as set forth in the appended claims.
- [0017]For a more complete understanding of the present invention, and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:
- [0018]
FIG. 1 is a diagram of an exemplary sigma-delta digital-to-analog converter (DAC); - [0019]
FIG. 2 is a diagram of a sigma-delta DAC used in an asymmetric digital subscriber line (ADSL) application; - [0020]
FIG. 3 is a diagram of a sigma-delta modulator with built-in reset and dithering logic, according to a preferred embodiment of the present invention; - [0021]
FIG. 4 is a diagram of a detailed view of the sigma-delta modulator shown inFIG. 3 , according to a preferred embodiment of the present invention; - [0022]
FIG. 5 is a pole-zero plot of a noise transfer function of the sigma-delta modulator shown inFIG. 3 , according to a preferred embodiment of the present invention; - [0023]
FIG. 6 is a data plot of a noise transfer function and signal transfer function of the sigma-delta modulator shown inFIG. 3 , according to a preferred embodiment of the present invention; - [0024]
FIG. 7 is a diagram of a detailed view of an implementation of the sigma-delta modulator shown inFIG. 3 , according to a preferred embodiment of the present invention; - [0025]
FIG. 8 is a flow diagram of an algorithm for determining if a sigma-delta modulator has become unstable, according to a preferred embodiment of the present invention; - [0026]
FIG. 9 is a data plot of continuous and quantized versions of an amplitude function used for dynamic dithering, according to a preferred embodiment of the present invention; - [0027]
FIG. 10 is a diagram of a dither generator, according to a preferred embodiment of the present invention; and - [0028]
FIG. 11 is a flow diagram of a process for maintaining sigma-delta modulator stability. - [0029]The making and using of the presently preferred embodiments are discussed in detail below. It should be appreciated, however, that the present invention provides many applicable inventive concepts that can be embodied in a wide variety of specific contexts. The specific embodiments discussed are merely illustrative of specific ways to make and use the invention, and do not limit the scope of the invention.
- [0030]The present invention will be described with respect to preferred embodiments in a specific context, namely a communications system that is adherent to several of the asymmetric digital subscriber line (ADSL) technical specifications, such as G.lite, ADSL, and ADSL+. The ADSL technical specifications can be found in a document entitled “International Telecommunication Union-ITU-T G.992.1-Series G: Transmission Systems and Media, Digital Systems and Networks, Digital Transmission Systems-Digital Sections and Digital Line System-Access Networks-Asymmetric Digital Subscriber Line (ADSL) Transceivers,” published June 1999. The invention may also be applied, however, to other asymmetric digital subscriber line systems, such as ADSL2 and ADSL2+, and to other digital subscriber line (DSL) systems. Note that in most instances, DSL and ADSL can be used interchangeably. Furthermore, the present invention can have applications in other systems wherein there is a need for a high frequency, high performance sigma-delta modulator.
- [0031]With reference now to
FIG. 1 , there is shown a diagram illustrating an exemplary sigma-delta digital-to-analog converter (DAC)**100**. The sigma-delta DAC**100**can be built out of a sigma-delta modulator**105**, a DAC**110**, and a filter**115**. The sigma-delta modulator**105**can have as an input, a band-limited signal, and can produce a digital output signal. The digital output signal can be a single bit signal or a multi-bit signal. The DAC**110**can be used to convert the digital output signal into a corresponding analog signal while the filter**115**can provide necessary smoothing and averaging. The filter**115**can also be used to help eliminate quantization noise that can reside in upper frequency ranges. Furthermore, there can also be a need to reject any images of the band-limited signal that may be present at the output Nyquist rate. - [0032]An advantage in using a sigma-delta modulator is its ability perform noise shaping. Noise shaping can be used to effectively push noise components to a frequency range that is above a frequency band of interest. The noise shaping ability of a sigma-delta modulator can be dependent upon its oversampling rate. Oversampling rates of approximately 64 times (and sometimes more) the clock rate of the signal being converted is considered to be adequate in moving the noise beyond the frequency band of interest. Unfortunately, this high oversampling rate can sometimes be a limiting factor in the decision to use a sigma-delta modulator, since in some applications, the high clock needed to operate the sigma-delta modulator may not be readily available.
- [0033]With reference now to
FIG. 2 , there is shown a diagram illustrating a sigma-delta DAC**200**that is being used in an asymmetric digital subscriber line (ADSL) application, such as G.lite, ADSL, ADSL+, and so on. Note that in many instances, ADSL and DSL can be used interchangeably. A DSL symbol, which is created in the frequency domain can be provided to an inverse Fourier transform unit**205**that can be used to create a time domain representation of the DSL symbol. The time domain version of the DSL symbol may then be provided to an interpolation chain**210**, which can be used to up-sample the time domain DSL symbol. The up-sampled time domain DSL symbol can then be provided to a sigma-delta modulator**215**wherein a digital output signal can be produced. The digital output signal may be a single bit signal or a multi-bit signal. The digital output can then be converted into an analog signal by a DAC and then provided to analog circuitry**220**where a filter can provide needed smoothing and averaging, along with noise elimination. Note that since the sigma-delta DAC comprises both the sigma-delta modulator**215**and the DAC (not shown), it may encompass both the sigma-delta modulator**215**and a portion of the analog circuitry**220**. The analog circuitry**220**may also contain circuits such as amplifiers and additional filters that may be needed to prepare the signal for transmission. - [0034]With reference now to
FIG. 3 , there is shown a diagram illustrating a sigma-delta modulator**300**with built-in reset and dithering logic, according to a preferred embodiment of the present invention. The sigma-delta modulator**300**can be used to create a sigma-delta DAC, such as the sigma-delta DAC**200**(FIG. 2 ), with the sigma-delta modulator**300**possibly being an implementation of the sigma-delta modulator functional block**215**shown inFIG. 2 . - [0035]The sigma-delta modulator
**300**, shown inFIG. 3 , can be logically divided into two paths: a lower path made up of two integrators (both labeled**310**) and an upper path made up of a dithering logic block**325**. A scale unit**305**can be present in the sigma-delta modulator**300**prior to the upper and lower paths. A summing point**315**can be used to combine signals from the lower and upper paths and a quantizer**320**can be used to convert a continuous-time signal into a discrete-time signal. The sigma-delta modulator**300**can also include a reset circuit**330**. - [0036]The lower path may be considered the main path of the sigma-delta modulator
**300**, wherein a discrete-time signal that is to be converted into an analog signal by the delta-sigma DAC**200**can be converted into a single bit (or multi-bit) signal. An initial scaling operation (via the scale unit**305**) on the discrete-time signal can be used to ensure a unity signal transfer function. The lower path, with its two integrators**310**may implement a sigma-delta modulator of order that is greater than or equal to two. Should the sigma-delta modulator be a first-order modulator, then a single integrator**310**may suffice. The summing point**315**combines the output of the dithering logic block**325**with the output of the two integrators**310**. A discussion of the dithering logic block**325**can be found below. Finally, the quantizer**320**may be used to quantize the combination of the outputs of the two integrators**310**with that of the dithering logic block**325**into one of two values (a single bit output) or one of 2^{N }values (an N-bit output). According to a preferred embodiment of the present invention, the sigma-delta modulator**300**produces a two-signal line output, implying that the quantizer**320**is a two-bit (four value) quantizer. - [0037]In DSL applications, when no data is being transmitted, the various DSL technical standards mandate that all zeros be transmitted. Unfortunately, the transmission of all zeros can present a problem for the sigma-delta modulator
**300**. The upper path, which comprises the dither logic block**325**can be used to help the sigma-delta modulator**300**circumvent problems associated with sending all zeros. This path also helps eliminate tones in the output for some input signals. - [0038]The DSL technical specifications, such as G.lite, ADSL, and ADSL+, can require a sigma-delta modulator
**300**to support signal bandwidths of 1.104 MHz and 2.208 MHz. A typical sigma-delta modulator using oversampling rates of 64 times oversampling (and more) may then need to operate frequencies of 282.624 MHz and greater. Such high operating frequencies may not be conducive low-cost, low-power applications, such as mass-market DSL modems. Since high operating frequencies may require higher technology fabrication lines, the cost of DSL modems using the high operating frequency sigma-delta modulator may be greater than one using a sigma-delta modulator operating at lower operating frequencies. - [0039]A sigma-delta modulator that is operating a lower oversampling rates may require that the modulator operate in a more aggressive manner. This can lead to a sigma-delta modulator that can be inherently unstable. To help mitigate the instability of such an aggressively designed sigma-delta modulator, the reset circuit
**330**can permit the sigma-delta modulator to recover. The reset circuit**330**can allow the sigma-delta modulator to reset itself (or be reset) once it has become unstable. Note that while the sigma-delta modulator**300**may be inherently unstable, the probability of the sigma-delta modulator**300**actually becoming unstable in normal operations remains low. - [0040]The discussion of the sigma-delta modulator
**300**presented below, uses a specific implementation of a sigma-delta modulator**300**that is used in a DSL modem and is capable of supporting DSL technical standards: G.lite, ADSL, and ADSL+. The sigma-delta modulator**300**operates at a clock frequency of 70.656 MHz and with signal bandwidths of up to 2.208 MHz. The specific values provided below are for this particular implementation of the sigma-delta modulator**300**. If the sigma-delta modulator**300**were to be used in a different application, with different operating frequencies, signal bandwidths, and so forth, it is expected that the values would be different. - [0041]With reference now to
FIG. 4 , there is shown a diagram illustrating a detailed view of the sigma-delta modulator**300**, according to a preferred embodiment of the present invention. An input signal, x′[n], can be provided to a signal input of the sigma-delta modulator**300**, wherein it can be provided to the scale unit**305**. According to a preferred embodiment of the present invention, the scale unit**305**can perform a multiplication of the input signal x′[n] with a first scale factor ‘sx’ and a second scale factor ‘stab’ wherein the first scale factor ‘sx’ can be used to scale the input signal x′[n] to ensure a unity signal transfer function (STF) for the sigma-delta modulator**300**. For the sigma-delta modulator**300**operating at a frequency of 70.656 MHz and for use in DSL applications, it is preferred that the scale factor ‘sx’ be essentially equal to 0.127 (0.12735753993847 to be precise). - [0042]The second scale factor ‘stab’ can be used to ensure that the output of the sigma-delta modulator
**300**remains within the stable output range of the sigma-delta modulator**300**, i.e., ensure that the magnitude of the output of the sigma-delta modulator**300**for a large variety of input signals remains below a certain value. For the sigma-delta modulator**300**, tests using a DC input have shown that a maximum stable output range to be around 0.34 (i.e., ‘stab’=0.34). However, for inputs that would be seen in a DSL system, such as single tones and discrete multi-tone (DMT) signals, the stability range can be significantly larger. For example, a DMT signal with a peak-to-average ratio (PAR) as large as 19 dB could be introduced at the input of the sigma-delta modulator**300**without having the sigma-delta modulator**300**becoming unstable. For input signals with a target PAR of 17 dB, this corresponds to a ‘stab’=0.79. Therefore, to be conservative, it is preferred that ‘stab’ be equal to 0.35. - [0043]Note that it may be possible to perform the scaling at lower frequencies (lower than the operating frequency of the sigma-delta modulator
**300**) by applying the scaling in earlier interpolation stages, prior to the sigma-delta modulator**300**(such as in the interpolation chain**210**(FIG. 2 )). - [0044]The input signal x′[n], after scaling, can be denoted x[n], may then be provided to both the upper and lower paths of the sigma-delta modulator
**300**. As discussed above, the lower path includes a pair of integrators**310**. In the sigma-delta modulator**300**, the integrators**310**are identical and the discussion of the first integrator can apply to the second integrator. The integrator**310**can comprise a pair of summing points**405**and**407**and a pair of first order integrators**409**and**411**, making the integrator**310**a second-order integrator. With two second order integrators, the sigma-delta modulator**300**is a fourth-order sigma-delta modulator. Note that in other designs, the integrators in the lower path may not necessarily be identical. - [0045]The first summing point
**405**can combine an input signal to the integrator**310**with a signal being fedback from the output of the integrator**310**with a signal being fedback from the output of the sigma-delta modulator**300**. The second summing point**407**can combine the output of the first integrator**409**with the signal being fedback from the output of the sigma-delta modulator**300**. According to a preferred embodiment of the present invention, the signals being fedback can be scaled prior to being combined in the first and the second summing points. For example, the signal being fedback from the output of the integrator**310**may be scaled (scale factor ‘a1’) prior to combining. The signal being fedback from the output of the sigma-delta modulator**300**may also be scaled (scale factors ‘b1’ and ‘b2’) prior to combining, but the scaling may be different for the first and the second summing points. It is preferred that the scale factor ‘a1’ be essentially equal to 0.0177 (0.01776642806738 to be precise), while the scale factor ‘b1’ be essentially equal to 0.109 (0.10950951822876 to be precise) and scale factor ‘b2’ be essentially equal to 0.419 (0.41941901226110 to be precise). Note that the feedback may be negative in nature, so the scaled feedback signals can be subtracted via the summing points**405**and**407**. - [0046]The second integrator
**310**may be similar to the first integrator**310**described above with the exception of the scale factors used in the fedback of the signals to the summing points**405**and**407**. The scale factors used in the second integrator**310**may be referred to with different labels from the scale factors used in the first integrator**310**to help reduce confusion. A scale factor ‘a2’ (analogous to the scale factor ‘a1’ from the first integrator**310**) can be preferred to be essentially equal to 0.000602 (0.00060236260760 to be precise), while scale factors ‘b3’ and ‘b4’ (analogous to the scale factors ‘b1’ and ‘b2’ from the first integrator**310**) can be preferred to be essentially equal to 1.003 and 0.852 (1.00399021176998 and 0.85229868030050 to be precise), respectively. Again, the feedback may be negative in nature, so the scaled feedback signals can be subtracted via the summing points**405**and**407**. - [0047]The upper path includes the dithering logic block
**325**, which can be used to help eliminate problems with idle tones that are generated for some low-amplitude input signals. The dithering logic block**325**implements a dynamic dithering algorithm rather than a static dithering algorithm to help avoid a reduction in the stability of the sigma-delta modulator**300**. Dithering can be applied to signals with magnitudes less than a specified threshold, while for signals with magnitudes greater than the specified threshold, no dither may be applied. Note that the amount of dithering may be graduated, depending upon the magnitude of the signal. For example, if the signal has a magnitude below a first threshold, then full dithering may be applied, while if the magnitude is between the first and a second threshold, then the dithering may be scaled down by a certain percentage. The dithering may be continually scaled until when the signal magnitude exceeds the specified threshold, no dithering is applied. - [0048]An input signal x[n], traversing the upper path, may be provided to a magnitude quantization unit
**415**. The magnitude quantization unit**415**can be used to scale the amount of dithering based upon the magnitude of the input signal x[n]. The magnitude quantization unit**415**comprises a magnitude unit**420**that can be used to obtain an absolute value measurement of the input signal x[n]. The magnitude of the input signal can then be provided to a non-linear quantizer**425**, which can quantize the magnitude of the input signal into one of several values. The non-linear quantizer**425**can be used to scale the amount of dithering applied to the input signal x[n] as described above. For example, if the magnitude of the input signal is less than the first threshold, then the non-linear quantizer**425**may quantize it to its maximum output value, while if the magnitude of the input signal is greater than the first threshold but less than the second threshold, then the non-linear quantizer**425**may quantize it to its second largest output value, and so forth. A delay**430**can be used to synchronize the output of the non-linear quantizer**425**with the output of the last integrator**411**in the lower path. The output of a dither generator**435**and the output of the non-linear quantizer**425**can then be combined by a summing point**440**, producing an output of the dithering logic block**325**. A detailed discussion of the dither generator**435**is provided below. - [0049]Signals from the upper and the low paths may be combined by the summing point
**315**, which can provide the combined signals to the quantizer**320**. The quantizer**320**is preferably a four (4) level quantizer (based upon the use of two output lines for the sigma-delta modulator**300**). The quantizer**320**may use four levels: −1, −1/3, 1/3, and 1. - [0050]With reference now to
FIG. 5 , there is shown a data plot illustrating a pole-zero plot of the noise transfer function (NTF) of the sigma-delta modulator**300**with scaling factors discussed above, according to a preferred embodiment of the present invention. The NTF for the sigma-delta modulator**300**features four poles and four zeroes. Note thatFIG. 5 displays a pole-zero plot with both quantized and unquantized poles**505**and zeroes**510**. Furthermore, the quantized and unquantized zeroes**510**remain in the same position on the unit circle, while some of the quantized and unquantized poles**505**move slightly (quantized poles**506**). The unquantized poles and zeroes are the result of representing the scale factors (discussed inFIG. 3 ) as double precision values and the quantized poles and zeroes are the result of rounding the representation of the scale factors. - [0051]With reference now to
FIG. 6 , there is shown a data plot illustrating plots of the NTF and the signal transfer function (STF) of the sigma-delta modulator**300**with scaling factors discussed above, as a function of frequency, according to a preferred embodiment of the present invention. A first curve**605**displays the NTF as a function of frequency and a second curve**610**displays the STF as a function of frequency.FIG. 6 displays both the quantized and unquantized NTF**605**and STF**610**for the sigma-delta modulator**300**. Note that both the quantized and unquantized NTF**605**and STF**610**lay essentially on top of one another, showing that any error resulting from the quantization is small and may be negligible. - [0052]With reference now to
FIG. 7 , there is shown a diagram illustrating a detailed view of the sigma-delta modulator**300**, wherein multiplications have been replaced with shifts and adds, according to a preferred embodiment of the present invention.FIG. 7 displays a possible implementation of the sigma-delta modulator**300**shown inFIG. 3 . As shown inFIG. 3 , the sigma-delta modulator**300**could require up to six (6) multiplications every clock cycle, wherein according to a preferred embodiment of the present invention, the clock is operating at 70.656 MHz. It can be possible to simplify the implementation by optimizing each of these six multiplications so that a multiplier is actually not required. For the feedback paths involving the scale factors ‘a1’ and ‘a2’ (feedback loops from the output of the integrator**310**to the input of the same integrator**310**), the scale factors can be represented in canonical signed digit (CSD) form and the multiplications can then be implemented as shifts and adds. The representation of numbers in their CSD form and using the CSD form to implement a multiplication as shifts and adds is considered well understood by those of ordinary skill in the art of the present invention. A Table #1 below lists the CSD forms of the scale factors ‘a1’ and ‘a2’ and a corresponding error from the use of the CSD forms.TABLE #1 CSD Representation of scale factors ‘a1’ and ‘a2’ Scale Factor Value CSD Form Percentage Error a1 0.01776642806738 2 ^{−6 }+ 2^{−9 }+ 2^{−12}0.3143 a2 0.00060236260760 2 ^{−11 }+ 2^{−13 }− 2^{−17}0.0597 - [0053]The four remaining multiplications involve an output of the quantizer
**320**as it is scaled (by scale factors ‘b1,’ ‘b2,’ ‘b3,’ and ‘b4’) and fedback to the integrators**310**. As discussed previously, the quantizer**320**is chosen to produce four levels: −1, −1/3, 1/3, and 1. A two-bit binary representation of the quantizer levels may be as follows: 10_{2 }can represent −1, 11_{2 }can represent −1/3, 00_{2 }can represent 1/3, and 01_{2 }can represent 1. For the scale factors ‘b1,’ ‘b2,’ ‘b3,’ and ‘b4’, two values for each scale factor can be stored in a lookup table (in CSD format): the scale factor itself and the scale factor multiplied by 1/3. The binary representation of the quantizer output can be used to determine if the scale factors are added or subtracted. A Table #2 below lists the CSD forms of the scale factors ‘b1,’ ‘b2,’ ‘b3,’ and ‘b4’ and a corresponding error from the use of the CSD forms after the scale factors have been divided by a peak reference level of 1.50, now referred to as “b1′,” “b2′,” “b3′,” and “b4′.”TABLE #2 CSD Representation of scale factors “b1′,” “b2′,” “b3′,” and “b4′” Percentage Scale Factor Value CSD Form Error b1′ 0.07300634548584 2 ^{−4 }+ 2^{−6 }− 2^{−8 }− 2^{−10}0.3230 b2′ 0.27961267484074 2 ^{−2 }+ 2^{−5 }− 2^{−9}−0.1129 b3′ 0.66932680784666 1 − 2 ^{−2 }− 2^{−4 }− 2^{−6}0.3807 b4′ 0.56819912020033 2 ^{−1 }+ 2^{−4 }+ 2^{−8}−0.3155 - [0054]Since the scale factors “b1′,” “b2′,” “b3′,” and “b4′” had been divided by the peak reference level of 1.50, the products, such as b1′*0.5=b1/3 and b1′*1.50=b1, can be computed with the same precision. A Table #3 below shows the resulting products, which can be stored in a lookup table.
TABLE #3 Stored product values Product Value Percentage Error b1 0.10950951822876 0.3230 b1/3 0.03650317274292 0.3230 b2 0.41941901226110 −0.1129 b2/3 0.13980633742037 −0.1129 b3 1.00399021176998 0.3807 b3/3 0.33466340392333 0.3807 b4 0.85229868030050 −0.3155 b4/3 0.28409956010017 −0.3155 - [0055]The implementation of the sigma-delta modulator
**300**shown inFIG. 7 makes use of the CSD representation of the scale factors to eliminate the need to perform multiplications.FIG. 7 displays the lower path of the sigma-delta modulator**300**, including the two integrators**310**, the summing point**315**(wherein the dithering produced in the upper path can be combined), and the quantizer**320**. Since the two integrators**310**are essentially identical, the discussion of one integrator**310**is sufficient. - [0056]The integrator
**310**comprises the pair of summing points**405**and**407**and the pair of first order integrators**409**and**411**. Due to the implementation, it may not be possible to clearly delineate the summing points**405**and**407**and the first-order integrators**409**and**411**. The summing point**405**comprises a series of two-input adders**705**,**707**, and**709**. The first two-input adder**705**can be used to combine the scaled output of the quantizer**320**with a delayed output of the first-order integrator**409**(via a delay**725**), while the second two-input adder**707**can be used to combine the output of the first two-input adder**705**with the input signal x[n]. The third two-input adder**709**can combine a rounded (via rounding circuit**711**) output of the second two-input adder**707**with a scaled output of the first integrator**310**. - [0057]However, the third two-input adder can also perform the summation (integration) performed by the first-order integrator
**409**. A limiting unit**715**can use logic to effectively place a limit on the precision of the output of the third two-input adder**709**. A second rounding circuit**717**can perform rounding on the output of the first-order integrator**409**. The scaling of the output of the quantizer**320**can be performed by a comparator, such as comparator**730**, which can be used to scale the quantizer output by the scale factor b1. According to a preferred embodiment of the present invention, depending upon the binary value of the output of the quantizer**320**, the comparator**730**can provide either b1, b1/3, −b1, or −b1/3 to an input of the first two-input adder**705**. - [0058]The reset circuit
**330**(not shown inFIG. 7 ) can be an automatic reset circuit that can reset the sigma-delta modulator**300**should the sigma-delta modulator**300**enter an unstable state. According to a preferred embodiment of the present invention, the second first-order integrator**411**of the second integrator**310**(hereby referred to as the fourth integrator) uses four (4) sign bits, thereby allowing values up to approximately 8.0 to be represented. The quantizer**320**, on the other hand, can be overloaded if its input exceeds an absolute value of 1.33. Simulations of typical DSL systems have shown that the sigma-delta modulator**300**can operate in the overloaded region, however, the typical maximum value is less than 2.0. Furthermore, simulations indicate that when the sigma-delta modulator**300**does become unstable, the fourth integrator will become saturated. Therefore, it can be sufficient to have the reset circuit**330**examine the two most significant bits of the output of the fourth integrator. Based upon the value of the two most significant bits of the output of the fourth integrator, the reset circuit**330**can perform a reset of the sigma-delta modulator**300**. - [0059]With reference now to
FIG. 8 , there is shown a flow diagram illustrating an algorithm**800**for determining if the sigma-delta modulator**300**has become unstable, according to a preferred embodiment of the present invention. The algorithm**800**may be executed by the reset circuit**330**in order to determine if the sigma-delta modulator**300**has become unstable and needs to be reset. The algorithm**800**may be implemented in hardware or as software. - [0060]At a specified frequency, perhaps once every clock cycle of the sigma-delta modulator
**300**, the reset circuit**330**can check the value of the two most significant bits of the fourth integrator (block**805**). The reset circuit**330**can check to see if the two bits are different from one another (block**810**). If the two bits are different, then the reset circuit**330**can increment an internal counter (block**815**). When the two bits are different, it is indicative of the fourth integrator operating in a saturated state. - [0061]If the two bits are not different (block
**810**), then the two bits must be equal. If the two bits are equal, then the reset circuit**330**can reset the internal counter, preferably to zero (block**820**). After either incrementing the internal counter (block**815**) or resetting the internal counter (block**820**), depending upon the value of the two bits, the reset circuit**330**can check the value of the internal counter to determine if it is equal to a specified value (block**825**). If the internal counter is equal to the specified value, then the fourth integrator has been operating in the saturated state for the specified value of consecutive samples, indicating that the sigma-delta modulator**300**has become unstable. The reset circuit**330**can then reset the sigma-delta modulator**300**(block**830**). The resetting of the sigma-delta modulator**300**may entail the setting of all integrator states back to zero (0) and the internal counter may be set back to zero (0). - [0062]In addition to using the reset circuit
**330**to detect the instability of the sigma-delta modulator**300**, it may be possible to have an external reset input for the sigma-delta modulator**300**, wherein an external device (such as a controller) or a user, can force the resetting of the sigma-delta modulator**300**. - [0063]With reference now to
FIG. 9 , there is shown a data plot illustrating continuous and quantized versions of an amplitude function used for dynamic dithering, according to a preferred embodiment of the present invention. The amplitude function can be expressed as:

(1*−abs*(*x*′)^{0.25})**d[n]*

wherein d[n] is the dithering sequence and x′=x/(sx*stab). Note that x′ is the input signal prior to the scale unit**305**and ‘sx’ and ‘stab’ are scale factors. The dithering sequence d[n] is a random variable with a triangular probability distribution function (TPDF) and can be formed by adding together two uniformly distributed random variables. - [0065]A first curve
**905**represents a continuous version of the amplitude function used for dynamic dithering and a second curve**910**represents a quantized version of the amplitude function. The quantization converts the amplitude of the amplitude function into several quantization levels, each of which is a power (negative) of two, to simplify implementation. Note that the quantization can be performed by the non-linear quantizer**425**(FIG. 4 ). According to a preferred embodiment of the present invention, if the amplitude function is less than {fraction (1/512)}, then the amplitude is quantized to 0.5, if the amplitude function is in the range ({fraction (1/512)}, 0.0625], then the amplitude is quantized to 0.25, if the amplitude function is in the range (0.0625, 0.3164], then the amplitude is quantized to 0.125, and if the amplitude function is greater than 0.3164, then the amplitude function is quantized to 0. - [0066]As discussed previously, the amount of dithering applied can vary depending upon the magnitude of the input signal. For example, if the magnitude of the input signal is less than 0.3164, then full dithering can be applied, and if the magnitude of the input signal is greater than 0.3164, then no dithering is applied.
- [0067]With reference now to
FIG. 10 , there is shown a diagram illustrating a detailed view of the dither generator**435**, according to a preferred embodiment of the present invention. The dither generator**435**can be based upon a pseudo-random number (PN) sequence generator. According to a preferred embodiment of the present invention, the dither generator**435**is based on the generating polynomial 1+x^{4}+x^{39}. The dither generator**435**comprises a sequentially coupled sequence of latches**1005**, wherein an output of one latch (such as latch**1010**) is provided to an input of the next latch in the sequence (such as latch**1015**). - [0068]The generating polynomial, 1+x
^{4}+x^{39}, can specify another set of connections for the outputs of the latches in the sequence of latches**1005**. The generating polynomial can be re-expressed as: x^{0}+x^{4}+x^{39}. The exponents of the ‘x’ terms specifies the output of corresponding latches being provided to an exclusive-or (XOR) gate**1025**. With the generating polynomial, 1+x^{4}+x^{39}, the outputs of latch #**0**(latch**1015**) and latch #**4**(latch**1020**) are provided to the XOR gate**1025**. The output of the XOR gate**1025**can then become the input of the last term of the generating polynomial, x^{39}, which is latch #**39**(latch**1030**). - [0069]With reference now to
FIG. 11 , there is shown a flow diagram illustrating a process**1100**that can be used in maintaining sigma-delta modulator stability, according to a preferred embodiment of the present invention. According to a preferred embodiment of the present invention, the process**1100**can be used to help maintain stability in the sigma-delta modulator**300**. A first operation (block**1105**) can be combining a dithering sequence, such as one produced in the dithering logic block**325**(FIG. 3 ), with an output produced by an integrator, such as the second integrator**310**(FIG. 3 ). A second operation in the process**1100**(block**1110**) can be to determine if an integrator is operating in saturation. According to a preferred embodiment of the present invention, the integrator that is being examined is the same integrator with the output being combined with the dithering sequence. If the integrator has been operating in saturation for a specified period of time, then the sigma-delta modulator**300**can be reset (block**1115**). The process**1100**can continue while the sigma-delta modulator**300**is in operation. - [0070]Although the present invention and its advantages have been described in detail, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
- [0071]Moreover, the scope of the present application is not intended to be limited to the particular embodiments of the process, machine, manufacture, composition of matter, means, methods and steps described in the specification. As one of ordinary skill in the art will readily appreciate from the disclosure of the present invention, processes, machines, manufacture, compositions of matter, means, methods, or steps, presently existing or later to be developed, that perform substantially the same function or achieve substantially the same result as the corresponding embodiments described herein may be utilized according to the present invention. Accordingly, the appended claims are intended to include within their scope such processes, machines, manufacture, compositions of matter, means, methods, or steps.

Patent Citations

Cited Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|

US5248972 * | Apr 3, 1992 | Sep 28, 1993 | Nokia Matkapuhelimet Oy | Delta-sigma analog-to-digital converter with overload dependent modulator order |

US5886586 * | Sep 5, 1997 | Mar 23, 1999 | The Regents Of The University Of California | General constant frequency pulse-width modulators |

US6061009 * | Mar 30, 1998 | May 9, 2000 | Silicon Laboratories, Inc. | Apparatus and method for resetting delta-sigma modulator state variables using feedback impedance |

US6157331 * | Oct 1, 1998 | Dec 5, 2000 | Tritech Microelectronics, Ltd. | Sigma delta modulator with automatic saturation detection and recovery |

US6166676 * | Oct 27, 1998 | Dec 26, 2000 | Sharp Kabushiki Kaisha | Correlating device and sliding correlating device using the same |

US6462685 * | Apr 5, 2001 | Oct 8, 2002 | Nokia Corporation | Dither signal insertion inversely proportional to signal level in delta-sigma modulators |

US6489907 * | Jul 30, 2001 | Dec 3, 2002 | Stmicroelectronics S.R.L. | Method of re-establishing the stability of a sigma-delta modulator and a circuit for implementing the method |

US6515540 * | Dec 10, 2001 | Feb 4, 2003 | Cirrus Logic, Inc. | High order multi-path operational amplifier with output saturation recovery |

US6621435 * | Oct 24, 2001 | Sep 16, 2003 | Stmicroelectronics S.R.L. | Method and circuit for improving the signal/noise ratio of a sigma-delta modulator |

US6765520 * | Apr 21, 2003 | Jul 20, 2004 | Texas Instruments Incorporated | Method and circuit for jamming digital filter while resetting delta sigma modulator |

US6768436 * | Apr 21, 2003 | Jul 27, 2004 | Texas Instruments Incorporated | Method and circuit resetting delta sigma modulator |

Referenced by

Citing Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|

US7358881 | Jul 22, 2005 | Apr 15, 2008 | Cirrus Logic, Inc. | Quantizer overload prevention for feed-back type delta-sigma modulators |

US7471223 * | Aug 18, 2006 | Dec 30, 2008 | Samsung Electronics Co., Ltd. | Delta-sigma modulator circuits in which DITHER is added to the quantization levels of methods of operating the same |

US7538701 * | Jun 9, 2007 | May 26, 2009 | Cosmic Circuits Private Limited | System and method for improving dynamic performance of a circuit |

US8472552 * | Dec 14, 2007 | Jun 25, 2013 | Icera, Inc. | Digital linear transmitter architecture |

US8620005 * | Jun 1, 2011 | Dec 31, 2013 | Suzhou Sonavox Eletronics Co., Ltd. | All-digital speaker system device |

US8766836 | Dec 14, 2012 | Jul 1, 2014 | Analog Devices, Inc. | Sigma delta modulator with dither |

US9184708 | Jul 30, 2012 | Nov 10, 2015 | Samsung Electronics Co., Ltd. | Audio signal processing method and audio signal processing apparatus therefor |

US20070018866 * | Jul 22, 2005 | Jan 25, 2007 | Melanson John L | Quantizer overload prevention for feed-back type delta-sigma modulators |

US20070040718 * | Aug 18, 2006 | Feb 22, 2007 | Samsung Electronics Co., Ltd. | Delta-sigma modulator circuits in which DITHER is added to the quantization levels of methods of operating the same |

US20070285297 * | Jun 9, 2007 | Dec 13, 2007 | Venkatesh T S | System and method for improving dynamic performance of a circuit |

US20100027711 * | Dec 14, 2007 | Feb 4, 2010 | Tajinder Manku | Digital linear transmitter architecture |

US20120099740 * | Jun 1, 2011 | Apr 26, 2012 | Suzhou Sonavox Eletronics Co., Ltd. | All-digital speaker system device |

EP2221976A2 * | Jul 13, 2006 | Aug 25, 2010 | Cirrus Logic, Inc. | Quantizer overload prevention for feed-back type delta-sigma modulators |

EP2737477A2 * | Jul 25, 2012 | Jun 4, 2014 | Samsung Electronics Co., Ltd. | Audio signal processing method and audio signal processing apparatus therefor |

EP2737477A4 * | Jul 25, 2012 | Apr 1, 2015 | Samsung Electronics Co Ltd | Audio signal processing method and audio signal processing apparatus therefor |

WO2007015736A1 * | Jul 13, 2006 | Feb 8, 2007 | Cirrus Logic, Inc. | Quantizer overload prevention for feed-back type delta-sigma modulators |

WO2013119945A1 * | Feb 8, 2013 | Aug 15, 2013 | Analog Devices, Inc. | Sigma delta modulator with dither |

Classifications

U.S. Classification | 341/143 |

International Classification | H03M3/00, H03M7/32 |

Cooperative Classification | H03M7/3035, H03M7/3011, H03M7/3028, H03M7/3008, H03M7/304 |

European Classification | H03M7/30B1C4, H03M7/30B1C2 |

Legal Events

Date | Code | Event | Description |
---|---|---|---|

Jun 10, 2004 | AS | Assignment | Owner name: TEXAS INSTRUMENTS INCORPORATED, TEXAS Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:ZOGAKIS, THOMAS NICHOLAS;KHASNIS, HIMAMSHU G.;VIJAYAVARDHAN, BALREDDY;REEL/FRAME:015462/0226;SIGNING DATES FROM 20040608 TO 20040609 |

Jun 22, 2009 | FPAY | Fee payment | Year of fee payment: 4 |

Mar 18, 2013 | FPAY | Fee payment | Year of fee payment: 8 |

Rotate